PNNL

Abstract

Sets of transmembrane (TM) helix-containing C-terminal protein regions from a single genome often show distinctive signature motifs adjacent to the helices. The combination of a characteristic motif, a TM region, and a cluster of basic residues at the extreme C-terminus forms a recognizable protein sorting and/or processing signal such as LPXTG and PEP-CTERM in bacteria and the CAAX box in eukaryotes. Hidden Markov models (HMMs) can detect and discriminate among distinct classes of protein C-terminal regions that share this tripartite architecture. This study reports the identification of multiple new prokaryotic C-terminal putative processing signals analogous to the previously described PEP-CTERM and LPXTG domains. One such signal, PGF-CTERM, occurs in at least twenty-nine archaeal species, with some genomes having more than fifty members. PGF-CTERM proteins include the major cell surface protein in Halobacterium, a glycoprotein with a partially characterized diphytanylglyceryl phosphate linkage in the C-terminal region. Comparative genomics identifies its cognate protein-sorting enzyme, a distant homolog to bacterial exosortases here designated archaeosortase, Additional newly defined C-terminal protein-sorting signals, mapped in silico to corresponding exosortase/archaeosortase-related machinery for sorting and post-translational modification, include a PXXXP-CTERM system operating in parallel to PGF-CTERM in two archaeal species, a PLPA-CTERM system in several alpha and deltaproteobacterial species, a PDSG-CTERM system unique to certain Verrucomicrobia, and a PEF-CTERM system in Methanosarcina mazei. Additional candidate protein sorting domains occur in lineages such as the Myxococcales, where some members lack homologs to either sortases or to exosortases, suggesting that several additional classes of C-terminal protein processing systems remain to be discovered.